Playing Becomes Performance: A Critique of PRT for Preschool Childen with Autism
Framing Fun as Fixing Quirks
In a recent study titled Improving Peer Interactions and Social Skills in Preschool Children With Autism Using Pivotal Response Treatment, lead author Yuxin Long and colleagues from Xihua University and affiliated rehabilitation centers in Sichuan and Guangzhou tested whether a short-term, play-based program called Pivotal Response Treatment (PRT) could improve the social behavior of autistic preschoolers. The team published their findings in the Journal of Autism and Developmental Disorders.
On its face, the study looks harmless, even encouraging. Children played imitation games, sang songs and matched cartoon cards. But beneath the colorful surface lies a deeper structural problem: the framing of autism itself. The paper is titled as a social skills success story. But read closely, and it's clear what's being measured isn't connection. It's compliance.
The authors are explicit about their goals: PRT, they say, is intended to "guide [children] back to a normal developmental trajectory." Autism is treated not as a neurotype, but as a deviation from an expected path. Improvement is defined not by wellbeing or self-direction, but by reductions in so-called "core autism symptoms." The underlying assumption is clear: autistic traits are errors to be corrected, not signals to be understood.
Target Behaviors, Trained Responses
It’s true that all preschoolers are taught social skills — but this isn’t that. In most early childhood classrooms, children are supported through modeling, guided play and emotional coaching that allows for individuality and autonomy. What PRT does instead is train children to perform specific behaviors on cue, using structured reinforcement to shape what counts as appropriate. That difference isn’t just technical. It’s ethical.
The study enrolled thirty children between the ages of four and six, all with mild to moderate autism diagnoses. Half were assigned to receive PRT, while the other half continued with conventional therapy. In this context, "conventional therapy" refers to standard rehabilitation practices commonly used in clinical settings, including speech therapy, sensory integration exercises and daily living skills training, typically guided by fixed routines and adult-led prompts rather than child-directed play. For the PRT group, sessions took place three times a week for eight weeks. During that time, children played games like "carrot cutting" and "cat and mouse" under the guidance of therapists who reinforced specific behaviors with praise or small rewards. These behaviors included initiating conversation, imitating movements, waiting their turn and sharing toys.
Each session had clearly defined "target behaviors" embedded in the game. When a child met a behavioral goal, they received reinforcement. When they didn’t, they were encouraged to try again. The goal was to increase "correct" behaviors over time, using structured incentives to shape social output. The study claims success: participants in the PRT group scored lower on autism symptom checklists and higher on scales measuring peer interaction.
But what those metrics actually capture is more troubling. They measure a child’s ability to mirror neurotypical norms. They reward performance, not authenticity. And they do it in the name of inclusion.
The Disappearance of Consent
The children in this study were four to six years old. They could not meaningfully consent to treatment, let alone understand its goals. This isn't a side detail. It's a structural concern. When interventions target identity-level traits in people too young to self-advocate, we aren't just helping them participate more fully. We are shaping them to be less inconvenient. The line between therapy and training gets blurry fast.
In fact, the study takes pride in the fact that PRT helps children demonstrate "non-targeted improvements." That is, even behaviors not explicitly taught in the session appeared to shift. This is framed as evidence of generalization. But it could just as easily be read as an early form of masking: the child learning not just what to do in the moment, but who to be in order to keep the rewards coming.
The long-term risks of this kind of training are well documented by autistic adults, many of whom describe behavioral interventions in childhood as experiences of coerced compliance. What looks like social improvement to a therapist may, years later, be remembered as erasure.
Who Gets to Define Progress?
The outcome measures in this study include the Childhood Autism Rating Scale (CARS), the Autism Behavior Checklist (ABC), and the Social Responsiveness Scale (SRS). Each of these tools assigns numeric scores to what it calls 'autism severity' or 'social skill level'—terms that seem neutral, but in practice reinforce a deep structural bias. Severity, in this context, usually means how far someone is from appearing neurotypical. Social skill level often translates to how well an autistic child can mimic neurotypical expectations. These scales don’t measure wellbeing.
What's being tracked is whether the child becomes easier to manage. More "social." Less "autistic." These tools treat deviation from neurotypical norms as dysfunction, then define success as the narrowing of that gap. It's the diagnostic equivalent of tidying up a room by hiding everything that doesn't match. None are based on autistic-defined goals.
Even more troubling, the people completing the assessments — therapists and parents — were not blinded to the group assignments. That means they knew which children received the intervention, and what kinds of changes the researchers hoped to see. The risk of expectation bias is high. So is the chance that progress was overestimated, especially given the short timeframe of the study.
Naturalistic in Form, Not in Ethics
PRT is often described as a "naturalistic developmental behavioral intervention" — a friendlier, more flexible form of ABA. It takes place in play settings, follows the child's interests and rewards effort as well as success. But these surface-level shifts do not change the underlying frame: autism as deficit, intervention as correction. Calling something 'naturalistic' doesn’t make it neutral. If the outcome is the same — suppression of difference — the setting doesn’t absolve the strategy.
In this study, PRT is pitched as low-cost, accessible and easy to implement in under-resourced settings. That may be true. But ethical harm does not become harmless just because it’s efficient. If the goal is to make autistic children easier to teach, more pleasant to interact with, and less visibly autistic — all without their informed input, without consent, and without any power to shape how they are seen — then the study isn’t measuring support. It's measuring normalization.
If an autistic child led the design of this study, would the same behaviors be labeled as progress — or protection? Every autism researcher needs to start asking this question before forming a hypothesis.
The Real Social Skill We’re Missing
No autistic people were involved in designing or interpreting this study. No autistic perspectives were cited. No alternative frameworks were tested. The research treats autism as something to work around, not something to understand from the inside.
This is not an oversight. It's a pattern.
The study's final paragraph calls for future research to use "objective behavioral measures" to validate findings. But the measures we actually need are not behavior, but wellbeing; not proximity to a norm, but depth of connection. What if we asked the children, years from now, how the intervention shaped their sense of self?
Help, For Whom?
There are gentler ways to support autistic kids. There are frameworks that prioritize emotional safety, communication diversity and sensory respect. That last one is often ignored, but it matters: sensory respect means recognizing and honoring the real, embodied experiences of autistic people — not forcing eye contact, not treating stimming as a problem, not designing classrooms that overwhelm, then blaming the child who melts down. There are tools for connection that do not rely on behaviorism.
But they require a different starting point. One where the child is not a problem to be solved, but a person to be heard and understood.
If this study proves anything, it’s that the desire to do good is not enough. Without autistic framing power, even play can become performance. Even games can become training. And even kindness can do harm, when the metric for success is who we stop being. If this is what “better” looks like, who is it better for?